---
title: "Constraints on Clitics in Amharic — Bliese 2026"
description: "This paper describes where Amharic clitics may and may not appear and asks what kind of grammar rule best captures each restriction."
authors: ["Loren F. Bliese"]
year: 2026
venue: "Folia Orientalia 20"
source: "https://wydawnictwo.pan.pl/index.php/fo/article/view/3306"
licence: "CC BY (journal licence per DOAJ; Folia Orientalia)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/bliese-2026-constraints-clitics-amharic/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/bliese-2026-constraints-clitics-amharic/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/3/) / Constraints on Clitics in Amharic

# Constraints on Clitics in Amharic

**Authors:** Loren F. Bliese · **Year:** 2026 · **Venue:** Folia Orientalia 20 · **Licence:** [CC BY (journal licence per DOAJ; Folia Orientalia)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://wydawnictwo.pan.pl/index.php/fo/article/view/3306) · [PDF](https://wydawnictwo.pan.pl/index.php/fo/article/download/3306/2525) · [OpenAlex](https://openalex.org/W7203567242)

**Citations:** 0 citations (openalex-snapshot · 2026-10-14)

## Summary
This paper describes where Amharic clitics may and may not appear and asks what kind of grammar rule best captures each restriction. It is a descriptive and theoretical study. The data are elicited example sentences that the author checked with native-speaker informants, and many of them are starred ungrammatical forms. It covers four areas:

- the coordinating clitics *inna*, *im* and *gin*
- pronominal clitics on infinitives
- subject–object suffix pairs on perfect-tense verbs
- subject marking in compound tenses formed with the auxiliaries *alla* and *nabbara*

The author identifies three, possibly four, constraints and proposes a formal device for each:

- ordered expansion rules
- ordering nominalization before cliticization
- selectional restrictions
- a deletion rule applied after phonology

Two explanatory threads run through the analysis. Some gaps exist because the forbidden form would be ambiguous. Exceptions to the "no double subject" rule appear where vowel assimilation would otherwise erase person or number contrasts. The work matters because it shows syntactic constraints that depend on surface phonology, and that are driven by the need to keep forms distinct.

## Key points
- *inna* ("and"), *im* ("also"/series-final "and") and *gin* ("but") all behave as enclitics. None may open a sentence, and none may attach to another of the three. To put *gin* first, speakers must host it on *daru* or *nagar*.
- Infinitives accept possessive (subject) clitics but not object clitics, so the object stays a free pronoun. The likely reason is that second-person subject and object clitics are identical, so allowing both would create ambiguity.
- On perfect verbs, a subject and object suffix cannot combine if both are first person, or both are second person. This holds across any gender or number. Reflexives instead use *ras* plus a possessive. The polite form and first person singular are the exceptions.
- The "no double subject constraint" works as follows:
  - In the present perfect (gerund + *alla*), only one of the two parts carries the subject marker.
  - The gerund's marker is lost only where *alla*'s initial vowel triggers assimilation.
  - This requires a deletion rule ordered after phonology.
- In the imperfect, most *alla* forms keep double subject marking. The author attributes this to opacity: deleting the suffix would merge forms such as "you (masc.)" and "she", or singular and plural.

## Interesting ideas
- A construction can be missing simply because filling it would make subject and object readings identical; the paper treats ambiguity avoidance as a real explanatory force. — For Mova (tokens): when the English morphology module faces an ambiguous form, it can prefer the reading that the grammar would actually allow, and record that preference as an explainable decision.
- Some well-formedness checks only work on the surface form produced after sound-level merging, not on the underlying structure. — For Mova (tokens): checks on English contractions and clitics (*'s*, *n't*, *'ll*) should run after normalization, so the tagger judges the forms that really occur.
- Certain function items are mutually exclusive, so a small list of forbidden pairings captures a constraint more simply than a general rule. — For Mova (tree): a short list of incompatible coordinators or discourse markers stacked on one head can flag suspicious `cc`/`advmod` attachments in parser output.
- Exceptions to a deletion rule can be predicted by asking whether applying it would collapse a person or number contrast. — For Mova (back): during generation and compression, block pronoun or agreement reductions whenever they would merge two distinct referents in the meaning graph.
- Ungrammatical variants were deliberately constructed and confirmed with informants, even when pauses were inserted between the clitics. — For Mova (learn): build starred minimal pairs as negative controls, so that evaluation verifies the system actually rejects impossible combinations.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

Fields: [Linguistics & languages](https://md.research.elamur.ai/free/en/shelves/linguistics/)

## Cite

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{bliese2026constraints,
  title = {{Constraints on Clitics in Amharic}},
  author = {Bliese, Loren F.},
  booktitle = {Folia Orientalia 20},
  year = {2026},
  url = {https://wydawnictwo.pan.pl/index.php/fo/article/view/3306}
}
```

- [← previous Consistency of LLMs to Comparative Statements in Mathematical Reasoning Tasks](https://md.research.elamur.ai/free/en/papers/san-2026-consistency-llms-comparative-statements-mathematical/)
- [next → Constructing a Japanese Claim Decomposition Dataset for Fact-Checking of LLM-Generated Texts](https://md.research.elamur.ai/free/en/papers/masano-2026-constructing-japanese-claim-decomposition-dataset/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/3/)